文献类型：期刊文献

中文题名：基于TM影像的农田防护林信息提取研究

英文题名：Research on Extracting Information of Shelterbelt Based on TM Remote Sensing Images

作者：邓荣鑫[1,2];李颖[1];张树文[1];王文娟[3]

第一作者：邓荣鑫

机构：[1]中国科学院东北地理与农业生态研究所;[2]中国科学院研究生院;[3]河南财经学院资源与环境科学系

第一机构：中国科学院东北地理与农业生态研究所,吉林长春130012

年份：2011

卷号：27

期号：1

起止页码：12-16

中文期刊名：农业系统科学与综合研究

外文期刊名：System Sciences and Comprehensive Studies in Agriculture

收录：CSCD:【CSCD2011_2012】；

基金：中国科学院知识创新工程重大项目(KZCX1-YW-08-02-01)

语种：中文

中文关键词：农田防护林;遥感;决策树;信息提取

外文关键词：shelterbelt; RS; decision tree algorithm; information extraction

摘要：农田防护林作为防护林的类型之一,在农田系统中具有重要作用。快速掌握农田防护林的动态变化,对于政府部门的管理和决策具有非常重要的意义。以吉林省德惠市区域作为试验区,选取2002年6月4日TM影像作为实验数据,结合研究区特点,将该区分为防护林地、其它林地、耕地、居民地、水域及未利用地等6类。通过光谱曲线分析、主成分变换分析和形状指数分析,利用决策树方法,最终提取出农田防护林信息。经验证,农田防护林的提取精度为85.52%,说明该方法可以为防护林的提取提供参考,为防护林的监测与管理提供数据支持。
As one kind of the protective plantations,shelterbelts play an important role in agricultural system.For government＇s management and decision,it＇s very important to fast obtain the dynamic change of shelterbelt.In this paper,Dehui City of Jilin Province is selected as a study area,the TM image acquired in 2002 is used as the study data.Based on the characteristic of the study area,land use in this area was divided into six classes—shelterbelt,other plantations,crops,residents,water and unused land.Through the spectral profile of original bands analysis,principal components analysis and the tasseled cap transform analysis,with the decision tree algorithm,the shelterbelt were extracted from the image.A result show that the accuracy is about 85.52%,Base on these analysises,this technique will have a consulting value in the extractions of the protective plantations,and provide technical support for the inspection and management of the protective plantations.